34 H. M. BERNARD. 



while externally the nuclei are for the most part functioning 

 as rod nuclei. It would thus be only a band of nuclei dowr 

 the middle which would be freest to travel towards tlio axis. 

 This fact gives us a clue to the origin of the zonal arrange- 

 ment of the retina into alternating nuclear and reticular layers. 

 It is fairly clear that if a band of nuclei travelled along 

 between two stationary layers such as the innermost nnd 

 outermost layers in an amphibian retina, and if, when they 

 started, they had cytoplasmic attachments to the limiting 

 membranes, they would almost certainly leave traces of those 

 attachments trailed along on each side of the stream, and the 

 accumulations of the trailings would separate them from 

 the stationary fringing layers. "We get, indeed, in this 

 somewhat startling and unexpected manner a perfectl}^ in- 

 telligible reason for the existence of the two reticular layers. 

 It is unexpected because, considering that all these cyto- 

 plasmic strands are living protoplasm, it would appear more 

 natural if they had readjusted themselves in the retina, letting 

 the nuclei pass on. The evidence, however, shows clearly 

 that this is not the case, and that they are to a large extent 

 trailed along and assist in the formation, at least, of the 

 inner reticular layer. I say " assist " because they appa- 

 rently only form its cytoplasmic basis ; other elements, as we 

 shall see later on, contribute to the final result. 



In studying thin sections of retinas of tadpoles I had 

 often been struck by the fact that from the extreme end of 

 the inner reticular layer irregular threads went off and 

 curved inwards towards the membrana limitans interna. This 

 is more striking in some cases than in others. Figs. 1 and 2 

 are sufiicientto show what is meant, but in some cases I have 

 seen it so marked that it looked as if the inner reticular 

 layer took its origin, at each end of the section, from 

 the internal limiting membrane, sometimes almost shutting 

 off the layer of " ganglionic cells " from those of the un- 

 differentiated rim. This appearance greatly puzzled me 

 until the discovery of the migration of the nuclei made 

 it clear that these threads which joined the inner reticular 



